Lgr5-Mediated Restraint of Β-Catenin Is Essential for B-Lymphopoiesis and Leukemia-Initiation

Lgr5-mediated restraint of β-catenin is essential for B-lymphopoiesis and leukemia-initiation

Kadriye Nehir Cosgun1, Mark E. Robinson1, Gauri Deb1, Xin Yang2, Gang Xiao1, Teresa Sadras1, Jaewoong Lee1, Lai N. Chan1, Kohei Kume1, Maurizio Mangolini3, Janet Winchester1, Zhengshan Chen1, Lu Yang1, Huimin Geng2, Shai Izraeli1,4, Joo Song5, Wing-Chung Chan5, Andrew G. Polson6, Hassan Jumaa7, Hans Clevers8 & Markus Müschen1

1Department of Systems Biology, City of Hope Comprehensive Cancer Center, Monrovia, CA 91016 2Department of Laboratory Medicine, University of California San Francisco, San Francisco, CA 94143 3Department of Haematology, Cambridge University, Cambridge, UK 4 Department of Human Molecular Genetics and Biochemistry, Sackler School of Medicine, Tel Aviv University, Tel Aviv, Israel 5Department of Pathology, City of Hope Comprehensive Cancer Center, Duarte, CA 6Genentech, Inc., South San Francisco, CA 94080 7Institute for Immunology, University of Ulm, Ulm, Germany 89081 8Hubrecht Institute, Royal Netherlands Academy of Arts and Sciences and University Medical Centre Utrecht, Utrecht, Netherlands

For correspondence: Markus Müschen, MD-PhD E-mail: [email protected] Phone: +1-626-218-5171 Department of Systems Biology, City of Hope Comprehensive Cancer Center, 1218 South Fifth Ave, Monrovia, CA 91016 bioRxiv preprint doi: https://doi.org/10.1101/2020.03.12.989277; this version posted March 13, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license. Upon productive immunoglobulin gene rearrangement, expression of a functional pre-B cell receptor (pre-BCR) initiates positive selection of pre-B cells, clonal expansion and self-renewal1-2. Studying mechanisms driving this first wave of B-lymphopoiesis, we identified the G-protein coupled receptor Lgr5 as an essential initiator of positive selection. Lgr5 was extensively studied as determinant of stem cell populations in multiple tissues3-6, but not in B-cells. While undetectable throughout the hematopoietic system, positively selected pre-B cells were marked with a sharp peak of Lgr5 expression. Conditional deletion of Lgr5 preceding the pre-BCR checkpoint induced negative selection and complete abortion of B-cell development. Proteomic studies of Lgr5-ablation revealed massive (>250-fold) accumulation of -catenin and suppression of MYC. Lgr5-deficient pre-B cells fully recovered by concurrent -catenin-deletion, demonstrating a central role of Lgr5-mediated restraint of -catenin at the pre-BCR checkpoint. In other cell types, -catenin/TCF4 complexes drive transcriptional activation of MYC7-9. Instead of TCF4, proximity-based interactome studies in pre-B cells identified the B- lymphoid transcription factors IKZF1 and IKZF310-11 as -catenin-binding partners, which had the opposite effect and caused transcriptional repression of MYC. On positively selected pre-B cells, Lgr5 prevented accumulation of -catenin and formation of complexes with IKZF1 and IKZF3, which relieved transcriptional repression of MYC. Activating β-catenin-mutations are common throughout all main types of cancer7-8, but were conspicuously absent in pre-B leukemia (B-ALL). Like pre-B cells, B- ALL cells were uniquely sensitive to genetic and pharmacological β-catenin hyperactivation, which recapitulated the effects of Lgr5-deletion and compromised colony formation and leukemia-initiation. A new LGR5 antibody-drug conjugate targeted leukemia-initiating cells in patient-derived B-ALL and achieved long-term disease-control. Likewise, small molecule hyperactivation of -catenin selectively killed B-ALL but not other cell types. Hence, Lgr5-mediated restraint of -catenin activation is essential for B-lymphopoiesis and revealed an unexpected vulnerability that can be leveraged for the treatment of drug-resistant B-ALL.

Lgr5 is essential for the initiation of B-lymphopoiesis Once early B-lymphocyte precursors have productively rearranged immunoglobulin (Ig) V, D and J gene segments, expression of a functional Ig -heavy chain as part of the pre-B cell receptor (pre-BCR) results in a strong positive selection signal to initiate clonal expansion and the first wave of B-lymphopoiesis1, 2. To study mechanisms of positive selection and pre-B cell self-renewal, we analyzed gene expression changes that are induced at the onset of Ig -heavy chain signaling. Among the most prominent changes at the pre-BCR checkpoint, we identified upregulation of the Leucine-rich repeat-containing G-protein coupled receptor 5 (Lgr5; Fig. 1a and b, Extended data figure 1a). While mRNA levels of Lgr5 were low or undetectable

A role of Lgr5 was not previously examined in B-lymphocyte development. To study the functional significance of sharp upregulation of Lgr5 at the pre-BCR checkpoint (Hardy Fractions C-D), we tested the consequences of inducible Cre-mediated deletion of Lgr5 in developing pre-B cells. Under cell culture conditions, inducible Lgr5-deletion caused rapid loss of competitive fitness of pre-B cells (Fig. 1d). In a conditional mouse model for deletion of Lgr5 from earliest stages of B-cell development (Mb1-Cre), pro-B cells (Hardy Fractions A-B) developed normally. However, B cell precursors could not progress past the pre- BCR checkpoint resulting in complete abortion of B-cell development in these mice (Fig. 1e-f, Extended data figure 2). Even deletion of only one allele of Lgr5 was sufficient to cause near-complete ablation of B-cell development, suggesting that even moderate reduction of Lgr5 gene expression may compromise positive selection at the pre-BCR checkpoint (Fig. 1e-f, Extended data figure 2). Interestingly, while deletion of Lgr5 compromised B-lymphopoiesis in vivo and competitive fitness in vitro, loss of Lgr5 did not significantly increase apoptosis and cell death (Extended data figure 9d, f). In addition, when deletion of Lgr5 was induced in mature B-cells after the pre-BCR checkpoint (Cd21-Cre; Extended data figure 3) or in germinal center B- cells (Aicda-Cre; Extended data figure 4), loss of Lgr5 only caused a modest reduction of marginal zone B- cells. All other mature B-cell subsets, including B1 cells remained unaffected. Lgr5-deficient germinal center (GC) B-cells underwent normal affinity maturation and were able to develop antigen-specific B cells in response to immunization (Extended data figure 4). Consistent with a sharp peak of gene expression in pre- B cells (Hardy Fractions C-D), these findings indicate that Lgr5 is uniquely required for positive selection at the pre-BCR checkpoint. Once mature B-cell populations have formed, they no longer depend on Lgr5 function for maintenance and affinity maturation in humoral immune responses. Lgr5 expression predicts poor clinical outcomes in patients with B-ALL While Lgr5 contributes to self-renewal of normal epithelial stem cells and tissue regeneration, this is also the case for cancer-initiating cells. For instance, lineage-tracing experiments revealed Lgr5 as determinant of tumor-initiation in intestinal adenomas21. In addition, Lgr5 marks colon cancer-initiating cells22, promotes metastasis23 and represents a therapeutic target for eradication of colon cancer stem cells (NCT02726334, 3

bioRxiv preprint doi: https://doi.org/10.1101/2020.03.12.989277; this version posted March 13, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license. NCT03526835)24, 25. For this reason, we studied LGR5 mRNA levels in patient-derived samples of B-cell acute lymphoblastic leukemia (B-ALL), the tumor that originates from pre-B cells, compared to common solid tumors and hematopoietic malignancies. While LGR5 is expressed in multiple solid tumors, including colon cancer, gastric, hepatocellular, breast cancer and glioma, we found the highest LGR5 mRNA levels in B-ALL samples (Fig. 1g). Consistent with confinement of LGR5 expression to the pre-B cell stage, LGR5 mRNA levels were high in most cases of B-ALL but low or undetectable in mature B-cell malignancies derived from naïve B-cells (chronic lymphocytic leukemia, mantle cell lymphoma), GC B-cells (Burkitt’s lymphoma, follicular lymphoma, diffuse large B-cell lymphoma) as well as post-GC B-cell malignancies (Hodgkin’s disease, multiple myeloma; Fig. 1g-h). Comparing LGR5 mRNA levels in cytogenetic B-ALL subtypes, we found high levels throughout all B-ALL subtypes, while MLL-rearranged B-ALL, mostly originating from pro- B cells, expressed LGR5 at lower levels (Extended data figure 1d). High expression levels of LGR5 in patient-derived B-ALL cells were confirmed at the protein level by Western blot and flow cytometry (Fig. 1i- j, Extended data figure 1e). To determine whether LGR5 expression may predict clinical outcomes in patients with lymphoid malignancies, LGR5 mRNA levels were measured at the time of diagnosis in one clinical trial for patients with B-ALL (P9906, n=207; Fig. 1k-m), six clinical cohorts of patients with mature B-cell malignancies and one T-cell lymphoma trial (Extended data figure 5). In each clinical cohort, patients were then segregated into two groups based on higher or lower than median LGR5 mRNA levels. Comparing persistence of minimal residual disease (MRD), overall survival and relapse-free survival between the two groups, higher LGR5 mRNA levels predicted poor clinical outcome in B-ALL (Fig. 1k-m) but not in any other hematological malignancies (Extended data figure 5). Mirroring Lgr5-functions during early B-cell development, these results suggest a disease-specific role of LGR5 in B-ALL. Lgr5 is essential for B-ALL leukemia-initiation To model common subtypes of B-ALL26, we transduced pre-B cells from Lgr5fl/fl mice12 with BCR-ABL1 or NRASG12D oncogenes and tamoxifen-inducible Cre (Cre-ERT2) or EV (ERT2; Fig. 2). Induction of Cre- mediated deletion of Lgr5 rapidly reduced competitive fitness of B-ALL cells under cell culture conditions (Fig. 2a-b, Extended data figure 6a-b) and abolished the ability of B-ALL cells to form colonies (Fig. 2c- d). When saturating numbers of full-blown B-ALL cells were injected into sublethally irradiated NSG mice, Cre-mediated deletion of Lgr5 substantially prolonged survival of transplant recipient mice (Extended data figure 6c-d). However, ultimately all mice died from fatal leukemia, suggesting that Lgr5 is not required for maintenance of established leukemia. In a serial transplant experiment based on limiting dilution of cell numbers (300, 6,000 and 120,000 injected B-ALL cells), Cre-mediated deletion of Lgr5 resulted in a profound defect of leukemia-initiation (Fig. 2e-g) and a ~40-fold reduction of leukemia-initiating cells (LIC; 1 in 1,031 to 1 in 40,055 cells), which suggests that Lgr5 is crucial for B-ALL leukemia-initiation. We next stained

bioRxiv preprint doi: https://doi.org/10.1101/2020.03.12.989277; this version posted March 13, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license. patient-derived B-ALL xenograft cells (PDX2) for LGR5 surface expression, sorted LGR5+ and LGR5- cells by flow cytometry and compared leukemia-initiation from 100,000, 10,000 and 1,000 injected cells (Extended data figure 7f, g). Compared to LGR5+ B-ALL cells, development of leukemia was delayed when LGR5- B- ALL cells were injected at low cell numbers, however both populations were able to initiate fatal leukemia. Interestingly, regardless of whether LGR5+ or LGR5- B-ALL cells were initially injected, B-ALL cells gave rise to mixed populations that contained similar fractions of LGR5+ and LGR5- B-ALL cells, (Extended data figure 7). Sorting and clonal outgrowth from single cells confirmed that LGR5- B-ALL cells can give rise to LGR5+ progeny and vice versa (Extended data figure 7d). These results suggest that LGR5 is required for LIC function in B-ALL but not a surface marker that would determine a permanent stem cell hierarchy. The single-cell sort experiment showed that most, if not all B-ALL cells have the capacity to express LGR5 and transiently acquire LIC-potential (Extended data figure 7d). Forced expression of LGR5 increased clonal competitiveness (Fig. 2h) and the ability to form colonies in patient-derived B-ALL xenografts (Extended data figure 7a) and murine B-ALL cells (Fig. 2i). A limiting-dilution transplantation experiment with patient- derived PDX2 B-ALL cells revealed that forced expression of LGR5 increased LIC-potential (Fig. 2j, Extended data figure 7b-c) and the frequency of LIC by ~8-fold (from 1 in 821 to 1 in 97; Fig. 2j). Stemness in pre-B cells and B-ALL is tied to positive selection and transient expression of LGR5 While stemness and LIC-capabilities in myeloid leukemia and solid tumors follow a developmental hierarchy27-29, these results suggest a transient model of stemness in B-ALL that is tied to positive selection and short-lived expression of LGR5, comparable to the narrow window of LGR5 expression at the pre-BCR checkpoint during normal B-cell development. Self-renewal in pre-B-cells is determined by positive selection based on their ability to express a functional pre-BCR, which then signals transient activation of Lgr5. A static, developmentally-determined hierarchy with a rare primitive stem cell at its apex, as in hematopoietic stem cells30, would broadly enable developmental progression but not positive selection of the few precursor cells that have productively rearranged Ig V, D and J segments to express a functional pre-BCR. Unlike myeloid and epithelial cells, we propose that self-renewal in early B-cell development is not a pre-determined fate but driven by positive selection signals from a functional pre-BCR. In B-ALL, self-renewal and positive selection are induced by oncogenic mimics of pre-BCR-signaling31-36. Hence, stemness and LIC-potential in B-ALL are tied to oncogenic signals that mimic positive selection signals from a functional pre-BCR. This scenario would reconcile the perplexing finding that unlike rare and phenotypically primitive stem cell populations in myeloid malignancies and solid tumors, LIC in B-ALL are very frequent and do not conform to a developmentally determined stem cell hierarchy37-40.

bioRxiv preprint doi: https://doi.org/10.1101/2020.03.12.989277; this version posted March 13, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license. Lgr5 functions as a negative regulator of -catenin To elucidate the mechanism of Lgr5-dependent leukemia-initiation and LIC survival in B-ALL, we assessed the consequences of Lgr5-deletion at the level of signal transduction. Lgr5 belongs to the rhodopsin-like class- A (subfamily 10) of G-protein coupled receptors, which signal through the G-protein-coupled receptor kinase 2 (GRK2) in B-cells41. Hence, we first performed a global proteomic analysis of phosphorylation changes upon inducible Lgr5-deletion in B-ALL cells. Consistent with a role of LGR5 in the regulation of WNT/-catenin- signaling, the most prominent phosphorylation-changes changes affected Axin 2, a negative regulator of - catenin42 and -catenin itself (S675; Fig. 3a). The phosphorylation on -catenin-S675 is mediated by cAMP- dependent protein kinase (PKA) and results in increased -catenin activity43, which was unexpected given the known role of LGR5 in positive regulation of WNT signaling in epithelial stem cells12. Other PKA-related molecules that were phosphorylated upon Lgr5-deletion included Akap12 and Pde8a, whereas phosphorylation of Myc and its target E2f2 was decreased (Fig. 3a). Measuring the impact of Lgr5-deletion at the transcriptional level, RNA-seq analysis revealed upregulation of multiple surface receptors that are associated with anergy and exhaustion of B- and T-cells, including Ctla4, Tigit, Prdm1, Ccr2, Dgka, Cd5, Il2ra (Cd25) and Cd244 (2B4). These changes were consistent with a recently identified signature of B-cell anergy and negative selection that depends on the Ikzf1 transcription factor44. The PKA regulatory subunit Prkar1b and PKA- scaffold Akap12 (Fig. 1i; Extended data figure 1d) were upregulated upon Lgr5-deletion, whereas Myc and E2f2 were downregulated (Fig. 3b). Western blot analyses of Lgr5-deletion in B-ALL cells confirmed upregulation of Dgka, Akap12, Cdkn2a, Prdm1 and Zap70 (Fig. 3c; Extended data figure 8a), while upregulation of anergy- and exhaustion markers (Cd5, Il2ra, Ccr2, Ctla4, Cd244) was confirmed by flow cytometry (Fig. 3i). Strikingly, Lgr5-deletion resulted in substantial depletion of Axin2 protein, which was paralleled by >250-fold accumulation of -catenin (Fig. 3c). Upregulation of Axin2 mRNA upon Lgr5- deletion is seemingly in contrast with loss of Axin2 protein. However, Axin2 represents a classical WNT/- catenin target gene45, hence massive accumulation of -catenin would be expected to increase transcriptional activation of Axin2. Consistent with loss of Axin2 and increased -catenin stability, phosphorylation of - catenin serine (S33, S37) and threonine (T41) residues was reduced (Fig. 3c). The sites in -catenin exon 3 are phosphorylated by GSK3 and initiate degradation. In addition, PKA-mediated -catenin phosphorylation of S675 identified by phosphoproteomic analyses (Fig. 3a) was confirmed by Western blot (Fig. 3 c). These results suggest a central role of LGR5 in negative regulation of WNT/-catenin signaling in pre-B cells and B-ALL.

Lgr5-dependent sequestration of Dishevelled2 enables GSK3-mediated negative regulation of -catenin To identify LGR5-interaction partners that could explain the underlying mechanism of LGR5 function and negative regulation of -catenin in B-ALL cells, we generated C-terminal fusion proteins between LGR5 and

-catenin-hyperactivation represents the functional equivalent of Lgr5-deletion If a main function of LGR5 was to limit -catenin activity, one would predict that inducible ablation of Lgr5 recapitulates the transcriptional program of hyperactive WNT/-catenin signaling. To test this hypothesis, we generated BCR-ABL1-driven B-ALL cells from mouse Ctnnb1ex3fl/+ bone marrow pre-B cells. Cre-mediated excision of exon 3 of -catenin in this mouse model removes GSK3 phosphorylation sites (S33, S37 and T41) for -catenin degradation (Harada et.al. 1999). Hence, activation of Cre in Ctnnb1ex3fl/+ B-ALL cells caused stabilization and accumulation of -catenin (Extended data figure 9). Gene expression changes resulting from -catenin hyperactivation mirrored multiple features of Lgr5-deletion including downregulation of Myc (Fig. 3e-f) and E2f2 (Fig. 3f) and induction of B-cell anergy phenotypes (Ctla4, Prdm1, Ccr2, Dgka, Cd5, Il2ra and Cd244; flow cytometry validation Fig. 3i-j). Gene set enrichment analysis revealed a striking global similarity of Ctnnb1ex3fl/+ and Lgr5fl/fl transcriptomes in B-ALL cells (Fig. 3g-h), suggesting that-catenin- hyperactivtion represents the functional equivalent of Lgr5-deletion at the transcriptional level.

Negative regulation of -catenin by Lgr5 is limited to early B-cell development Studying the entire spectrum of hematopoietic populations, Lgr5 expression was found exclusively at the pre- BCR checkpoint during early B-cell development (Fig. 1a-c; Extended data figure 1b-c). This extremely narrow window of gene expression was mirrored by the specific requirement of Lgr5 at the pre-BCR checkpoint: deletion of Lgr5 in mature and GC-B-cells had only very subtle effects (Extended data figure 3- 4) whereas deletion prior to the pre-BCR checkpoint resulted in complete loss of B-cell production (Fig. 1f, Extended data figure 2). Strikingly, deletion of Lgr5 in pre-B cells caused a >250-fold increase of -catenin protein levels (Fig. 3c, Fig. 4g) but had no measurable effect on -catenin levels in mature B-cells (Fig. 4g). Consistent with the identification of negative regulation of -catenin as central mechanism of Lgr5 function, these results corroborate that mature B-cells no longer depend on Lgr5 function (Fig. 4h).

Negative regulation of -catenin in B-cell malignancies Oncogenic WNT/-catenin signaling and activating mutations that increase stability and transcriptional activity of -catenin represent a common feature throughout all main types of cancer7, 8. In a comprehensive analysis of 66,820 samples encompassing 15 types of cancer (COSMIC)52, we identified 4,971 activating

Lgr5 function in B-ALL compared to colon cancer cells While not interchangeable, Wnt ligands (e.g. Wnt3a) of FZD receptors and R-spondin ligands (e.g. Rspo1) of Lgr5 cooperate to potentiate Wnt/β-catenin signaling. Rspo1 is thought to act as ligand for Lgr512, 18, which then counteracts the RNF43 and ZNRF3 E3 ubiquitin ligases to stabilize Fzd receptors for prolonged Wnt/β- catenin signaling. Based on recent evidence, however, Rspo2 can directly inhibit RNF43 and ZNRF3 and promote Wnt/β-catenin signaling even in the absence of Lgr556. In B-ALL cells, deletion of Lgr5 resulted in a >250-fold accumulation of nuclear -catenin, however, Wnt3a and Rspo1 had no measurable effect on - catenin levels, neither alone nor in combination (Extended data figure 11b). Since deletion of Lgr5 caused massive accumulation of nuclear -catenin and loss of self-renewal capacity in B-ALL cells, we tested whether LGR5-deletion in human colon cancer cells (LOVO) had similar effects. Deletion of LGR5 in LOVO cells was

-catenin forms complexes with Ikzf1 and Ikzf3 in pre-B cells for transcriptional repression of Myc Decreased phosphorylation of E2f2 (S314) and Myc (T79; Fig. 3a) and transcriptional downregulation of E2f2 and Myc was identified as a common feature of Lgr5-deletion (Fig. 3b) and direct -catenin hyperactivation (Fig. 3f). Consistent with a major role of Myc downstream of Lgr5 and -catenin, we found that Myc transcriptional targets were globally suppressed both upon deletion of Lgr5 (Lgr5fl/fl; Fig. 6a) and -catenin hyperactivation (Ctnnb1ex3fl/+; Fig. 6b). For this reason, we tested whether reconstitution of Myc could rescue the deleterious effects of Lgr5 deletion and -catenin hyperactivation. Myc overexpression restored competitive fitness of B-ALL cells that was lost upon Lgr5-deletion (Extended data figure 14a, c) or -

bioRxiv preprint doi: https://doi.org/10.1101/2020.03.12.989277; this version posted March 13, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license. catenin hyperactivation (Extended data figure 14b, d). Indeed, both Lgr5-deletion and inducible stabilization of -catenin caused profound suppression of Myc in B-ALL cells (Fig. 3e; Extended data figure 14), which is in striking contrast to Myc as a classical target of -catenin-mediated transcriptional activation9. To elucidate why -catenin functions as transcriptional repressor of Myc in pre-B cells, unlike other cell types, we identified -catenin interacting proteins in B-ALL cells. To this end, we transformed pre-B cells from Mb1Cre-ERT2 x Lgr5fl/fl x Ctnnb1flag/+ knockin mice that express FLAG-tagged -catenin from one allele. Upon Cre-mediated deletion of Lgr5, we identified binding partners by pull-down of FLAG-tagged -catenin and proteomic analyses (Fig 6c; Extended data figure 15a). Surprisingly, the proteins with the highest nuclear enrichment of binding to -catenin included Ikzf1 (Ikaros) and Ikzf3 (Aiolos), two pre-B cell-specific transcription factors that function as transcriptional repressors of Myc10. The interaction between -catenin and Ikzf1 was confirmed by co-immunoprecipitation and Western blot (Fig. 6c). We propose that that -catenin binds to the MYC promoter in pre-B cells, as in other cell types. Unlike -catenin/TCF4 complexes in other cell types7, 9, however, -catenin associates in pre-B cells with Ikzf1 and Ikzf3. Given that both Ikzf1 and Ikzf3 are specifically active at the pre-BCR checkpoint10, 11, the discovery of nuclear complexes between -catenin and Ikzf1-Ikzf3 in pre-B cells could explain that -catenin functions as transcriptional repressor of Myc in pre-B and B-ALL (Fig. 6e) as opposed to transcriptional activation in other cell types. Besides negative regulation of Myc, -catenin-Ikzf1 complexes more broadly promoted an Ikzf1-dependent transcriptional program (Extended data figure 15b-c), that recapitulates multiple features of Ikzf1-dependent transcriptional programs of B-cell anergy and negative selection44. Therapeutic targeting of Lgr5+ leukemia-initiating cells (LIC) in B-ALL Given that Lgr5 serves a critical role in self-renewal and survival of LIC in B-ALL (Fig. 2a-j), we tested whether eradication of Lgr5+ cells may be useful for therapeutic purposes. To eradicate cancer stem cells in colon cancer, an antibody-drug conjugate (ADC) targeting LGR5 was recently developed and demonstrated potent efficacy and safety in vivo24. Since Lgr5 mRNA levels in B-ALL cells are similar or even higher compared to colon cancer (Fig. 1g), we tested the efficacy of LGR5-ADC in a xenotransplantation model based on B-ALL PDX. NSG mice bearing B-ALL PDX were treated with anti-LGR5 (8E11) conjugated to a potent microtubule inhibitor, monomethyl auristatin E (MMAE) or an isotype control antibody (anti-gD) linked to MMAE. Compared to anti-gD-vc-MMAE, anti-LGR5-vc-MMAE substantially reduced leukemia burden and significantly extended overall survival (P=0.028; Extended data figure 16a-b). However, when treatment was stopped (day 50), mice recipients ultimately relapsed and died from B-ALL. These findings suggest that repeated injections with anti-LGR5-vc-MMAE achieved long-term disease control but not complete LIC-eradication.

Pharmacological hyperactivation of -catenin leverages a unique vulnerability in B-ALL cells Our mechanistic experiments revealed that hyperactivation of -catenin and -catenin-mediated transcriptional repression of MYC represent the underlying basis of the unexpected dependency of B-ALL cells on LGR5. Hence, as an alternative to LGR5-ADC mediated targeting of LGR5+ B-ALL cells, we tested pharmacological hyperactivation of -catenin, and, presumably, -catenin-mediated repression of MYC, to eradicate LIC in patient-derived B-ALL. To this end, we tested two FDA-approved small molecule inhibitors of GSK3 (9- ING-41 and LY2090314; Extended data figure 16) that have demonstrated safety in clinical trials for patients with metastatic cancer and acute myeloid leukemia (AML)58. Inhibition of GSK3 would be expected to increase stability of -catenin in a similar way as deletion of the GSK3-phosphorylation sites (exon 3) in the Ctnnb1ex3fl/+ mouse model (Fig. 3e, Extended data figure 9a)59. Strikingly, LY2090314 had profound effects on B-ALL cells at low nanomolar concentrations but not colon cancer or myeloid leukemia cells (Fig. 6f-g). Even at >40-fold higher concentrations, LY2090314 had no substantial toxic effects on colon cancer and myeloid leukemia cells (Fig. 6g). The toxic effect on B-ALL cells was paralleled by massive accumulation of -catenin, comparable to baseline levels in colon cancer cells (Fig. 6h, l). While structurally similar to LY2090314 (Extended data figure 16e), 9-ING-41 was not selective for B-ALL and killed B-ALL, colon cancer and myeloid leukemia cells only at much higher concentrations. Unlike LY2090314, 9-ING-41 failed to increase -catenin protein levels in B-ALL. In colon cancer cells, -catenin levels were >250-fold higher than in B-ALL cells, but remained unchanged after treatment with 9-ING-41. To test our hypothesis that accumulation of -catenin represents the mechanistic basis of LY2090314-mediated toxicity in B-ALL cells, we deleted CTNNB1 in the human B-ALL cell line BV173 using Cas9-RNPs and screening of clones for CTNNB1-deletion from single cells. Interestingly, deletion of CTNNB1 conferred complete resistance of B- ALL cells to LY2090314. Importantly, preventing -catenin upregulation in response to LY2090314- treatment also mitigated suppression of Myc, which would otherwise result from LY2090314-induced - catenin hyperactivation (Fig. 6i-k). Further corroborating -catenin-induced repression of Myc as the underlying mechanism of LY2090314-induced toxicity in B-ALL cells, we tested the effect of LY2090314 on -catenin and Myc levels in resistant colon cancer and myeloid leukemia cells as well as sensitive B-ALL cells. -catenin levels were constitutively high in colon cancer and did not increase upon drug-treatment. Resistant myeloid leukemia cells responded to LY2090314 by strong upregulation of -catenin (Fig. 6l). However, unlike B-ALL cells, massive accumulation of -catenin did not come at the expense of Myc, presumably because myeloid cells lack expression of IKZF1 and IKZF3.

IKZF1 and IKZF3 set the threshold for -catenin-induced transcriptional repression of MYC

bioRxiv preprint doi: https://doi.org/10.1101/2020.03.12.989277; this version posted March 13, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license. These pre-B cell-specific transcription factors function as repressors of MYC10 and associate with -catenin in B-ALL cells (Fig. 6e, 6l; Extended data figure 15). Likewise, colon cancer cells, lacking expression of IKZF1 and IKZF3, expressed -catenin at >250-fold higher baseline levels than B-ALL cells, but without any apparent impact on expression levels of MYC. -catenin/TCF4 complexes drive transcriptional activation of MYC7, 9 in many cell types, including colon cancer and CML cells. However, -catenin associated with the B- lymphoid transcription factors Ikzf1 and Ikzf3 has the opposite function in pre-B cells and functions as a powerful transcriptional repressor of Myc. Hence, Ikzf1 and Ikzf3 compete with TCF4 for binding to -catenin and thereby determine the outcome of -catenin signaling, e.g. transcriptional activation (TCF4) or repression (Ikzf1, Ikzf3) of MYC. Importantly, transcriptional repression of MYC by IKZF1 and IKZF3 is restricted to pre-B cells10, which explains the narrow window of LGR5 expression and sensitivity to -catenin hyperactivation at the pre-BCR checkpoint. For instance, deletion of Lgr5 and -catenin hyperactivation did not affect mature B-cells (Fig. 4g-h; Extended data figure 3, 4). IKZF1 and IKZF3 are essential for oncogenic signaling and therapeutic targets in multiple myeloma60, a B-lymphoid malignancy derived from terminally differentiated B-cells. In contrast, both IKZF1 and IKZF361 are tumor suppressors and frequently deleted in human B-ALL. Importantly, multiple oncogenic drivers in B-ALL, including TCF3-PBX162 and BCR-ABL163 can activate -catenin for increased proliferation and survival. Hence, deletion of IKZF1 or IKZF3 may allow for increased activation of -catenin in B-ALL cells, while LGR5 functions as feedback regulator to buffer fluctuations of -catenin signaling. Importantly, IKZF1-deletion represents a particularly impactful predictor of poor clinical outcomes in patients with B-ALL64. However, patients with IKZF1-deletion retain IKZF3 function and deletion of the two transcription factors is mutually exclusive61. In our analysis, both MXP5 and BV173 B-ALL cells harbored IKZF1 deletions. However, despite IKZF1-deletion, B-ALL remained fully sensitive to GSK3-inhibition (Fig. 6f; Extended data figure 14e) and downregulated MYC expression upon treatment with LY2090314 (Fig. 6j, 6l). While these responses were slightly diminished compared to B-ALL samples without IKZF1-deletion, these results suggest that IKZF3 alone is sufficient to repress MYC and induce cell death. Hence, acute hyperactivation of -catenin (e.g. by GSK3-inhibition, LY2090314) likely represents a powerful therapeutic approach even for high-risk B-ALL cases with IKZF1-deletion.

Concluding remarks Previous studies demonstrated that hyperactivation of -catenin can have deleterious consequences in the hematopoietic system65, 66 and that multiple mechanisms, e.g. LMBR1L-mediated ubiquitination of FZD6 and LRP667 are necessary to achieve optimal dosage and prevent toxic levels of -catenin hyperactivation. Here we show that LGR5-mediated sequestration of DVL2 is essential to curb -catenin activation at the pre-BCR checkpoint. Pre-B cells that fail to express a functional pre-BCR do not upregulate Lgr5. In the absence of

Acknowledgments: We would like to thank Dr. Michael Kahn, Department of Molecular Medicine, COHCCC, for critical discussions, Lars Klemm and Dr. Anna Hecht for their help with technical aspects of experiments and current and former members of the Müschen laboratory for their support and helpful discussions. Research in the Müschen laboratory is funded by the NIH through an NCI Outstanding Investigator Award R35CA197628 (to M.M.), U10CA180827 (to M.M.), R01CA137060, R01CA157644, R01CA172558 and R01CA213138 (to M.M.), the Howard Hughes Medical Institute HHMI-55108547 (to M.M.), the Norman and Sadie Lee Foundation (to M.M.), and the Falk Trust through a Falk Medical Research Trust Catalyst Award (to M.M.), the Pediatric Cancer Research Foundation (PCRF), and the California Institute for Regenerative Medicine (CIRM) through DISC2-10061. M.M. is a Howard Hughes Medical Institute (HHMI) Faculty Scholar. Conflicts of interest: A.G.P. is an employee with Genentech, Inc., South San Francisco, CA, and contributed to the development of Lgr5-ADC. None of the other co-authors have any conflicting interests.

References: 1. Rajewsky K. Clonal selection and learning in the antibody system. Nature. 381: 751-758 (1996). 2. R. Hardy, K. Hayakawa. B cell development pathways. Annu. Rev. Immunol. 19, 595-621 (2001). 3. T. Yasuda, H. Sanjo, G. Pagès, Y. Kawano, H. Karasuyama, J. Pouysségur, M. Ogata, T. Kurosaki. Erk kinases link pre-B cell receptor signaling to transcriptional events required for early B cell expansion. Immunity 4, 499-508 (2008). 4. N. Barker, J. van Es, J. Kuipers, P. Kujala, M. van den Born, M. Cozijnsen, A. Haegebarth, J. Korving, H. Begthel, P.J. Peters, H. Clevers. Identification of stem cells in small intestine and colon by marker gene Lgr5. Nature 449, 1003–1007 (2007). 5. M. Huch, C. Dorrell, S. Boj, J. H. van Es, V. S. Li, M. van de Wetering, T. Sato, K. Hamer, N. Sasaki, M. J. Finegold, A. Haft, R. G. Vries, M. Grompe, H. Clevers. In vitro expansion of single Lgr5+ liver stem cells induced by Wnt-driven regeneration. Nature 494, 247–250 (2013). 6. S. Buczacki, H. Zecchini, A. Nicholson, R. Russell, L. Vermeulen, R. Kemp, D. J. Winton. Intestinal label- retaining cells are secretory precursors expressing Lgr5, Nature 495, 65–69 (2013). 7. P. J. Morin, A. B. Sparks, V. Korinek, N. Barker, H. Clevers, B. Vogelstein, K. W. Kinzler. Activation of beta-catenin-Tcf signaling in colon cancer by mutations in -catenin or APC. Science 275, 1787-90 (1997). 8. B. Rubinfeld, P. Robbins, M. El-Gamil, I. Albert, E. Porfiri, P. Polakis. Stabilization of beta-catenin by genetic defects in melanoma cell lines. Science, 275, 1790-2 (1997). 14

bioRxiv preprint doi: https://doi.org/10.1101/2020.03.12.989277; this version posted March 13, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license. 9. T. C. He, A. B. Sparks, C. Rago, H. Hermeking, L. Zawel, L. T. da Costa, P. J. Morin, B. Vogelstein, K. W. Kinzler. Identification of c-MYC as a target of the APC pathway. Science, 281, 1509-12 (1998). 10. S. Ma, S. Pathak, M. Mandal, L. Trinh, M. R. Clark, R. Lu. Ikaros and Aiolos inhibit pre-B-cell proliferation by directly suppressing c-Myc expression. Mol. Cell Biol. 17, 4149-58 (2010). 11. T. A. Schwickert, H. Tagoh, S. Gültekin, A. Dakic, E. Axelsson, M. Minnich, A. Ebert, B. Werner, M. Roth, L. Cimmino, R. A. Dickins, J. Zuber, M. Jaritz, M. Busslinger. Stage-specific control of early B cell development by the transcription factor Ikaros. Nat. Immunol. 3, 283-93 (2014). 12. W. de Lau, N. Barker, T. Low, B. K. Koo, V. S. Li, H. Teunissen, P. Kujala, A. Haegebarth, P. J. Peters, M. van de Wetering, D. E. Stange, J. E. van Es, D. Guardavaccaro, R. B. Schasfoort, Y. Mohri, K. Nishimori, S. Mohammed, A. J. Heck, H. Clevers. Lgr5 homologues associate with Wnt receptors and mediate R-spondin signalling. Nature 476, 293–297 (2011). 13. S. Angers, C. J. Thorpe, T. L. Biechele, S. J. Goldenberg, N. Zheng, M. J. MacCoss, R. T. Moon. The KLHL12-Cullin-3 ubiquitin ligase negatively regulates the Wnt-beta-catenin pathway by targeting Dishevelled for degradation. Nat. Cell Biol. 8, 348-357 (2006). 14. W. Chen, D. ten Berge, J. Brown, Ahn S, L. A. Hu, W. E. Miller, M. G. Caron, L. S. Barak, R. Nusse, R. J. Lefkowitz. Dishevelled 2 recruits beta-arrestin 2 to mediate Wnt5A-stimulated endocytosis of Frizzled 4. Science 301, 1391-4 (2003). 15. A. Yu, J. F. Rual, K. Tamai, Y. Harada, M. Vidal, X. He, T. Kirchhausen. Association of Dishevelled with the clathrin AP-2 adaptor is required for Frizzled endocytosis and planar cell polarity signaling. Dev. Cell 1,129-41 (2007). 16. X. Huang, J. C. McGann, B. Y. Liu, R. N. Hannoush, J. R. Lill, V. Pham, K. Newton, M. Kakunda, J. Liu, C. Yu, S. G. Hymowitz, J. A. Hongo, A. Wynshaw-Boris, P. Polakis, R. M. Harland, V. M. Dixit. Phosphorylation of Dishevelled by protein kinase RIPK4 regulates Wnt signaling. Science 339, 1441-5 (2013) 17. V. Bryja, D. Gradl, A. Schambony, E. Arenas, G. Schulte. Beta-arrestin is a necessary component of Wnt/beta-catenin signaling in vitro and in vivo. Proc. Natl. Acad. Sci. U S A 104, 6690-5 (2007). 18. K. S. Carmon, X. Gong, Q. Lin, A. Thomas, Q. Liu. R-spondins function as ligands of the orphan receptors LGR4 and LGR5 to regulate Wnt/β-catenin signaling. Proc. Natl. Acad. Sci. USA 108, 11452–11457 (2011) 19. M. Fereshteh, T. Ito, J. J. Kovacs , C. Zhao, H. Y. Kwon, V. Tornini, T. Konuma, M. Chen, R. J. Lefkowitz, T. Reya. β-Arrestin2 mediates the initiation and progression of myeloid leukemia. Proc. Nat.l Acad. Sci. U S A 109, 12532-7 (2012). 20. C. Bonnans, M. Flacelière, F. Grillet, C. Dantec, J. P. Desvignes, J. Pannequin, D. Severac, E. Dubois, F. Bibeau, V. Escriou, P. Crespy, L. Journot, F. Hollande, D. Joubert. Essential requirement for β-arrestin2 in mouse intestinal tumors with elevated Wnt signaling. Proc. Natl. Acad. Sci. U S A 109, 3047-52 (2012). 21. A. G. Schepers, H. J. Snippert, D. E. Stange, M. van den Born, J. H. van Es, M. van de Wetering, H. Clevers. Lineage tracing reveals Lgr5+ stem cell activity in mouse intestinal adenomas. Science 337, 730– 5 (2012). 22. M. Shimokawa, Y. Ohta, S. Nishikori, M. Matano, A. Takano, M. Fujii, S. Date, S. Sugimoto, T. Kanai, T. Sato. Visualization and targeting of LGR5+ human colon cancer stem cells. Nature 545, 187-192 (2017). 23. F. de Sousa e Melo, A. V. Kurtova, J. M. Harnoss, N. Kljavin, J. D. Hoeck, J. Hung, J. E. Anderson, E. E. Storm, Z. Modrusan, H. Koeppen, G. J. Dijkgraaf, R. Piskol, F. J. de Sauvage. A distinct role for Lgr5+ stem cells in primary and metastatic colon cancer. Nature 543, 676-680 (2017). 15

bioRxiv preprint doi: https://doi.org/10.1101/2020.03.12.989277; this version posted March 13, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license. 24. M. Junttila, M. Weiguang, X. Wang, B. Wang, T. Pham, J. Flygare, S. Yu, S. Yee, D. Goldenberg, C. Fields, J. Eastham-Anderson, M. Singh, R. Vij, J. Hongo, R. Firestein, M. Schutten, K. Flagella, P. Polakis and A. G. Polson. Targeting LGR5+ cells with an antibody-drug conjugate for the treatment of colon cancer. Sci. Transl. Med. 7, 314ra186 (2015). 25. Two antibodies against LGR5, BNC101 (NCT02726334) and MCLA-158 (NCT03526835) were FDA- approved and demonstrated safety and efficacy in clinical trials for patients with gastrointestinal cancers. 26. J. Zhang, C. G. Mullighan, R. C. Harvey, G. Wu, X. Chen, M. Edmonson, K. H. Buetow, W. L. Carroll, I. M. Chen, M. Devidas, D. S. Gerhard, M. L. Loh, G. H. Reaman, M. V. Relling, B. M. Camitta, W. P. Bowman, M. A. Smith, C. L. Willman, J. R. Downing, S. P. Hunger. Key pathways are frequently mutated in high-risk childhood acute lymphoblastic leukemia: a report from the Children's Oncology Group. Blood 118, 3080-7 (2011). 27. T. Lapidot, C. Sirard, J. Vormoor, B. Murdoch, T. Hoang, J. Caceres-Cortes, M. Minden, B. Paterson, M. A. Caligiuri, J. E. Dick. A cell initiating human acute myeloid leukaemia after transplantation into SCID mice. Nature 367, 645-8 (1994). 28. E. Passegué, C. H. Jamieson, L. E. Ailles, I. L. Weissman. Normal and leukemic hematopoiesis: Are leukemias a reacquisition of stem cell characteristics. Proc. Natl. Acad. Sci. USA 100, 11842-9 (2003). 29. D. Bonnet, J. E. Dick. Human acute myeloid leukemia is organized as a hierarchy that originates from a primitive hematopoietic cell. Nat. Med. 3, 730-737 (1997). 30. T. Reya, A. W. Duncan, L. Ailles, J. Domen, D. C. Scherer, K. Willert, L. Hintz, R. Nusse, I. L. Weissman. A role for Wnt signalling in self-renewal of haematopoietic stem cells. Nature 423, 409-14 (2003). 31. A. C. Shaw, W. Swat, R. Ferrini , L. Davidson, F. W. Alt. Activated Ras signals developmental progression of recombinase-activating gene (RAG)-deficient pro-B lymphocytes. J. Exp. Med. 189, 123-9 (1999). 32. T. Tretter, A. E. Ross, D. I. Dordai, S. Desiderio. Mimicry of pre-B cell receptor signaling by activation of the tyrosine kinase Blk. J. Exp. Med. 198, 1863-73 (2003). 33. F. Klein, N. Feldhahn, L. Harder, H. Wang, M. Wartenberg, W. K. Hofmann, P. Wernet, R. Siebert, M. Müschen. The BCR-ABL1 kinase bypasses selection for the expression of a pre-B cell receptor in pre-B acute lymphoblastic leukemia cells. J. Exp. Med.199, 673-85 (2004). 34. N. Feldhahn, F. Klein, J. L. Mooster, P. Hadweh, M. Sprangers, M. Wartenberg, M. M. Bekhite, W. K. Hofmann, S. Herzog, H. Jumaa, J. D. Rowley, M. Müschen. Mimicry of a constitutively active pre-B cell receptor in acute lymphoblastic leukemia cells. J. Exp. Med. 201, 1837-52 (2005). 35. L. J. Anderson, R. Longnecker. EBV LMP2A provides a surrogate pre-B cell receptor signal through constitutive activation of the ERK/MAPK pathway. J. Gen. Virol. 89, 1563-8 (2005). 36. Z. Chen, S. Shojaee, M. Buchner, H. Geng , J. W. Lee, L. Klemm, B. Titz, T. G. Graeber, E. Park, Y. X. Tan, A. Satterthwaite, E. Paietta, S. P. Hunger, C. L. Willman, A. Melnick , M. L. Loh, J. U. Jung, J. E. Coligan, S. Bolland, T. W. Mak, A. Limnander, H. Jumaa, M. Reth, A. Weiss, C. A. Lowell, M. Müschen. Signalling thresholds and negative B-cell selection in acute lymphoblastic leukaemia. Nature 521, 357-61 (2015). 37. P. N. Kelly, A. Dakic, J. M. Adams, S. L. Nutt, A. Strasser. Tumor growth need not be driven by cancer stem cells. Science 317, 337 (2007). 38. C. le Viseur, M. Hotfilder, S. Bomken, K. Wilson, S. Röttgers, A. Schrauder, A. Rosemann, J. Irving, R. W. Stam, L. D. Shultz, J. Harbott, H. Jürgens, M. Schrappe, R. Pieters, J. Vormoor. In childhood acute lymphoblastic leukemia, blasts at different stages of immunophenotypic maturation have stem cell properties. Cancer Cell 14, 47-58 (2008).

bioRxiv preprint doi: https://doi.org/10.1101/2020.03.12.989277; this version posted March 13, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license. 39. K. Rehe, K. Wilson, S. Bomken, D. Williamson, J. Irving, M. L. den Boer, M. Stanulla, M. Schrappe, A. G. Hall, O. Heidenreich, J. Vormoor. Acute B lymphoblastic leukaemia-propagating cells are present at high frequency in diverse lymphoblast populations. EMBO Mol. Med. 5, 38-51 (2013). 40. Y. Aoki, T. Watanabe, Y. Saito, Y. Kuroki, A. Hijikata, M. Takagi, D. Tomizawa, M. Eguchi, M. Eguchi- Ishimae, A. Kaneko, R. Ono, K. Sato, N. Suzuki, S. Fujiki, K. Koh, E. Ishii, L. D. Shultz, O. Ohara, S. Mizutani, F. Ishikawa. Identification of CD34+ and CD34- leukemia-initiating cells in MLL-rearranged human acute lymphoblastic leukemia. Blood 125, 967-80 (2015). 41. T. I. Arnon, Y. Xu, C. Lo, T. Pham, J. An, S. Coughlin, G. W. Dorn, J. G. Cyster. GRK2-dependent S1PR1 desensitization is required for lymphocytes to overcome their attraction to blood. Science 333, 1898-903 (2011). 42. J. Behrens, B. A. Jerchow, M. Würtele, J. Grimm, C. Asbrand, R. Wirtz, M. Kühl, D. Wedlich, W. Birchmeier. Functional interaction of an axin homolog, conductin, with beta-catenin, APC, and GSK3beta. Science 280, 596-9 (1998). 43. S. Taurin, N. Sandbo, Y. Qin, D. Browning, N. O. Dulin. Phosphorylation of beta-catenin by cyclic AMP- dependent protein kinase. J. Biol. Chem. 281, 9971-6 (2006). 44. T. A. Schwickert, H. Tagoh, K. Schindler, M. Fischer, M. Jaritz, M. Busslinger. Ikaros prevents autoimmunity by controlling anergy and Toll-like receptor signaling in B cells. Nat Immunol. 11, 1517- 1529 (2019). 45. B. Wang, L. Zhao, M. Fish, C. Y. Logan, R. Nusse. Self-renewing diploid Axin2(+) cells fuel homeostatic renewal of the liver. Nature 524, 180-5 (2015). 46. D. V. Tauriello, I. Jordens, K. Kirchner, J. W. Slootstra, T. Kruitwagen, B. A. Bouwman, M. Noutsou, S. G. Rüdiger, K. Schwamborn, A. Schambony, M. M.Maurice. Wnt/β-catenin signaling requires interaction of the Dishevelled DEP domain and C terminus with a discontinuous motif in Frizzled. Proc. Natl. Acad. Sci. U S A Apr 3 109, E812-20 (2012). 47. X. Zeng, H. Huang, K. Tamai, X. Zhang, Y. Harada, C. Yokota, K. Almeida, J. Wang, B. Doble, J. Woodgett, A. Wynshaw-Boris, J. C. Hsieh, X. He. Initiation of Wnt signaling: control of Wnt coreceptor Lrp6 phosphorylation/activation via frizzled, dishevelled and axin functions. Development 135, 367-75 (2008). 48. V. Rauschenberger, D. B. Bernkopf, S. Krenn, K. Jalal, J. Heller, J. Behrens, M. Gentzel, A. Schambony. The phosphatase Pgam5 antagonizes Wnt/β-Catenin signaling in embryonic anterior-posterior axis patterning. Development 144, 2234-2247 (2017). 49. V. Brault, R. Moore, S. Kutsch, M. Ishibashi, D. H. Rowitch, A. P. McMahon, L. Sommer, O. Boussadia, R. Kemler. Inactivation of the beta-catenin gene by Wnt1-Cre-mediated deletion results in dramatic brain malformation and failure of craniofacial development. Development 128, 1253-64 (2001). 50. J. Huelsken, R. Vogel, B. Erdmann, G. Cotsarelis, W. Birchmeier. Beta-catenin controls hair follicle morphogenesis and stem cell differentiation in the skin. Cell 105, 533-45 (2001). 51. M. Cobas, A. Wilson, B. Ernst, S. J. Mancini, H. R. MacDonald, R. Kemler, F. Radtke. Beta-catenin is dispensable for hematopoiesis and lymphopoiesis. J. Exp. Med. 199, 221-9 (2004). 52. https://cancer.sanger.ac.uk/cosmic/gene/analysis?ln=CTNNB1 53. https://www.proteinatlas.org/ENSG00000168036-CTNNB1/pathology 54. https://portals.broadinstitute.org/ccle/data 55. K. S. Yan, C. Y. Janda, J. Chang, G. X. Y. Zheng, K. A. Larkin, V. C. Luca, L. A. Chia, A. T. Mah, A. Han, J. M. Terry, A. Ootani, K. Roelf, M. Lee, J. Yuan, X. Li, C. R. Bolen, J. Wilhelmy, P. S. Davies, H. 17

bioRxiv preprint doi: https://doi.org/10.1101/2020.03.12.989277; this version posted March 13, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license. Ueno, R. J. von Furstenberg, P. Belgrader, S. B. Ziraldo, H. Ordonez, S. J. Henning, M. H. Wong, M. P. Snyder, I. L. Weissman, A. J. Hsueh, T. S. Mikkelsen, K. C. Garcia, C. J. Kuo. Non-equivalence of Wnt and R-spondin ligands during Lgr5+ intestinal stem-cell self-renewal. Nature 545, 238-242 (2017). 56. E. Szenker-Ravi, U. Altunoglu, M. Leushacke, C. Bosso-Lefèvre, M. Khatoo, H. Thi Tran, T. Naert, R. Noelanders, A. Hajamohideen, C. Beneteau, S. B. de Sousa, B. Karaman, X. Latypova, S. Başaran, E. B. Uyguner, K. Vleminckx, N. Barker, H. Kayserili, B. Reversade. RSPO2 inhibition of RNF43 and ZNRF3 governs limb development independently of LGR4/5/6. Nature 557, 564-569 (2018). 57. R. Habas, Y. Kato, X. He. Wnt/Frizzled activation of Rho regulates vertebrate gastrulation and requires a novel Formin homology protein Daam1. Cell 107, 843-54 (2001). 58. Both 9-ING-41 (NCT04218071, NCT03678883) and LY2090314 (NCT01287520, NCT01632306, NCT01214603) are FDA-approved small molecule inhibitors of GSK3 and are in clinical trial for patients with metastatic cancer and acute myeloid leukemia (AML). 59. C. Liu, Y. Li, M. Semenov, C. Han, G. H. Baeg, Y. Tan, Z. Zhang, X. Lin, X. He. Control of beta-catenin phosphorylation/degradation by a dual-kinase mechanism. Cell, 108, 837-47 (2002). 60. Krönke J, Udeshi ND, Narla A, Grauman P, Hurst SN, McConkey M, Svinkina T, Heckl D, Comer E, Li X, Ciarlo C, Hartman E, Munshi N, Schenone M, Schreiber SL, Carr SA, Ebert BL. Lenalidomide causes selective degradation of IKZF1 and IKZF3 in multiple myeloma cells. Science, 343, 301-305 (2014). 61. Mullighan CG, Goorha S, Radtke I, Miller CB, Coustan-Smith E, Dalton JD, Girtman K, Mathew S, Ma J, Pounds SB, Su X, Pui CH, Relling MV, Evans WE, Shurtleff SA, Downing JR. Genome-wide analysis of genetic alterations in acute lymphoblastic leukaemia. Nature, 446, 758-764 (2007). 62. McWhirter JR, Neuteboom ST, Wancewicz EV, Monia BP, Downing JR, Murre C. Oncogenic homeodomain transcription factor E2A-Pbx1 activates a novel WNT gene in pre-B acute lymphoblastoid leukemia. Proc Natl Acad Sci U S A. 96: 11464-11469 (1999). 63. Coluccia AM, Vacca A, Duñach M, Mologni L, Redaelli S, Bustos VH, Benati D, Pinna LA, Gambacorti- Passerini C. Bcr-Abl stabilizes beta-catenin in chronic myeloid leukemia through its tyrosine phosphorylation. EMBO J. 26: 1456-1466 (2007). 64. Mullighan CG, Su X, Zhang J, Radtke I, Phillips LA, Miller CB, Ma J, Liu W, Cheng C, Schulman BA, Harvey RC, Relling MV, Shurtleff SA, Campana D, Borowitz MJ, Pui CH, Smith M, Hunger SP, Willman CL, Downing JR. Children's Oncology Group. Deletion of IKZF1 and prognosis in acute lymphoblastic leukemia. N Engl J Med. 360: 470-480 (2009). 65. Luis TC, Naber BA, Roozen PP, Brugman MH, de Haas EF, Ghazvini M, Fibbe WE, van Dongen JJ, Fodde R, Staal FJ. Canonical wnt signaling regulates hematopoiesis in a dosage-dependent fashion. Cell Stem Cell. 9: 345-356 (2011). 66. Scheller M, Huelsken J, Rosenbauer F, Taketo MM, Birchmeier W, Tenen DG, Leutz A. Hematopoietic stem cell and multilineage defects generated by constitutive beta-catenin activation. Nat Immunol. 7: 1037- 1047 (2006) 67. Choi JH, Zhong X, McAlpine W, Liao TC, Zhang D, Fang B, Russell J, Ludwig S, Nair-Gill E, Zhang Z, Wang KW, Misawa T, Zhan X, Choi M, Wang T, Li X, Tang M, Sun Q, Yu L, Murray AR, Moresco EMY, Beutler B. LMBR1L regulates lymphopoiesis through Wnt/β-catenin signaling. Science. 364 pii: eaau0812 (2019). 68. T. Le Carrour, S. Assou, S. Tondeur, L. Lhermitte, N. Lamb, T. Reme, V. Pantesco, S. Hamamah, B. Klein, J. De Vos. Amazonia: An Online Resource to Google and Visualize Public Human whole Genome Expression Data. The Open Bioinformatics Journal 4, 5-10 (2010). 69. E. C. Townsend, M. A. Murakami, A. Christodoulou et al., The Public Repository of Xenografts Enables Discovery and Randomized Phase II-like Trials in Mice. Cancer Cell 29, 574-586 (2016).

bioRxiv preprint doi: https://doi.org/10.1101/2020.03.12.989277; this version posted March 13, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license. Figure legends Figure 1: Identification of Lgr5 as an essential requirement for B-lymphopoiesis (a) Microarray analyses (GSE53896) of Lgr5 mRNA levels in FACS-sorted B220+ CD19+ pre-B cell fractions from VH81x transgenic mice with functional (Blnk+/+, n=4) or non-functional (Blnk-/-, n=4) pre-BCR signaling chains are shown (left). Non-VH81x transgenic Rag1-/- pro-B cells were used as controls (n=3). RNA-seq analysis of Blnk-/- pre-B-cells that were reconstituted with 4-hydroxy-tamoxifen (4-OHT) inducible Blnk for inducible activation of pre-BCR signaling (right). Lgr5 mRNA levels at 0 to 5 hours after 4-OHT addition (GSE73562) were indicated. (b) Rag2-/- pro-B cells were transduced with a 4-OHT-inducible Ig -heavy chain (HC) to initiate pre-BCR signaling. Lgr5 surface levels were measured by flow cytometry before and 24 hours after induction (n=9). (c) ImmGen microarray data sets (GSE15907, GSE37448) were analyzed for Lgr5 mRNA levels in 18 murine hematopoietic populations including 10 B-cell subsets. Pre-B cell stages (Fractions C-D) are highlighted in red. (d) IL7-dependent pre-B cells from the bone marrow of Lgr5fl/fl mice were transduced with GFP-tagged, 4-OHT-inducible Cre (Cre-ERT2) or empty vector controls (ERT2). Following 4- OHT induction, percentages of GFP+ cells, normalized to fraction of GFP+ cells before induction, were monitored by flow cytometry (n=3). Immature B cells (CD21- CD23-), marginal zone B cells (CD21+ CD23- ) and follicular B cells (CD21+ CD23+) were analyzed by flow cytometry of splenic tissue from Mb1Cre/+ Lgr5+/+, Mb1Cre/+ Lgr5fl/+, Mb1Cre/+ Lgr5fl/fl mice (n=3). (e) Relative frequencies and (f) representative FACS stainings are shown. Microarray data for LGR5 mRNA levels were studied in hematological malignancies and solid tumors. In (g) LGR5 mRNA levels are shown for 7 hematological malignancies and 5 solid tumor types (red dots) as well as their normal counterpart or cell of origin (gray dots)68. In (h), LGR5 mRNA levels in B- ALL samples (red) were compared to other hematological malignancies (gray)69. (i) Western blot analyses were performed to measure protein levels of LGR5 and AKAP12 in a panel of patient derived xenografts (PDX) and Jurkat T-ALL cells. (j) Surface levels of LGR5 expression was determined by flow cytometry in a panel human B-ALL, B-cell lymphoma, multiple myeloma, T-ALL, peripheral T-cell lymphoma (PTCL) and myeloid leukemia cell lines and PDX samples. (k) Minimal residual disease (MRD) status from patients with pediatric high-risk B-ALL (COG P9906) was determined by flow cytometry on day 29 post treatment. LGR5 expression levels were compared in patients with MRD+ versus MRD- status (P=0.025). (l) B-ALL patients from the same clinical cohort were segregated into two groups based on whether LGR5 mRNA levels were higher (LGR5high) or lower (LGR5low) than median expression. Relapse-free survival (RFS) was assessed in the two groups by Kaplan-Meier analysis (log-rank test, P=0.008). (m) This cohort included a subset of Ph- like ALL (n=43) patients. Overall survival (OS) was assessed in the LGR5high and LGR5low groups by Kaplan- Meier analysis, however the difference in OS did not reach statistical significance (log-rank test, P=0.07).

(d) To identify LGR5-interaction partners, we generated C-terminal fusion proteins between the cytoplasmic tail of LGR5 and the bacterial biotin-ligase BirA. Proximity-based analyses of proteins that were biotinylated by LGR5-BirA fusions (Bio-ID) identified dishevelled2 (DVL2), its scaffold -arrestin (ARRB2), as well as the DVL2 clathrin adapter AP2 (AAK1, AP2A2, AP2B1) as central interaction partners. Bio-ID analysis were performed for human PDX2 (x-axis) and SUPB15 (y-axis) B-ALL cells. WNT/-catenin-related proteins are labeled in red, PKA-related proteins in blue. (e) Lgr5fl/fl and Ctnnb1ex3fl/+ B-ALL cells and Lgr5fl/fl B-ALL cells carrying retroviral Cas9 for deletion of Ctnnb1 (Lgr5fl/fl sg-Ctnnb1) were transduced with Cre (for deletion of Lgr5 or stabilization of -catenin) or EV. Western blot analysis was performed to measure - catenin and Myc protein levels, using -actin as loading control. For Cas9-mediated deletion of Ctnnb1, results for two sg-Ctnnb1 guides, 1A12 and 2F4, are shown. (f) RNA-seq analysis of Ctnnb1ex3fl/+ BCR-ABL1 B- ALL cells carrying ERT2 or Cre-ERT2 was performed 1 day after 4-OHT addition, gene expression changes fl/fl ex3fl/+ shown as heatmap. Gene expression changes [log2-fold] in Lgr5 (x-axis) and Ctnnb1 (y-axis) B-ALL cells were compared in a scatter plot (g) and through GSEA analysis (h). Common gene expression changes in Lgr5fl/fl (i) and Ctnnb1ex3fl/+ (j) B-ALL cells that affected cell surface receptors were validated by flow cytometry. FACS dot plots for double stainings for Cd19 with Cd5, Il2ra (Cd25), Ccr2, Cxcr4, Ctla4 and Cd244 (2B4) are shown without (EV) or 3 days after (Cre) 4-OHT-mediated induction of Cre-ERT2. Numbers in FACS plots denote mean fluorescence intensities.

(a) Bone marrow pre-B cells from Ctnnb1ex3fl/+ mice were transformed by BCR-ABL1 and transduced with GFP-tagged Cre-ERT2 or ERT2 constructs. Changes of percentages of GFP+ cells were monitored for 12 days upon 4-OHT addition (n=3). (b) 10,000 Ctnnb1ex3fl/+ BCR-ABL1 B-ALL cells carrying Cre-ERT2 or ERT2 were plated for colony formation assays 2 days after 4-OHT treatment. Representative images and colony numbers are shown at 10 days after plating (n=3). (c) Cell cycle phases of Ctnnb1ex3fl/+ BCR-ABL1 B-ALL cells carrying Cre-ERT2 or ERT2 were measured by EdU incorporation in combination with DAPI staining. (d) Lgr5fl/fl BCR-ABL1 B-ALL cells were transduced with GFP-tagged Cre-ERT2 or ERT2 and transduced with Cas9 and guides for non-targeting controls (sg-NTC) or guides targeting Ctnnb1 (sg-Ctnnb1). Deletion of Ctnnb1 was confirmed by Western blot (d) in clonal cell lines that grew out from single clones (Extended data figure 10). Competitive fitness of Lgr5fl/fl B-ALL clones was measured based on deletion of Lgr5 (Cre- ERT2) and/or deletion of Ctnnb1 (sg-Ctnnb1). (e) Lgr5fl/fl B-ALL cells with (Cre-ERT2) or without (ERT2) deletion of Lgr5 in combination with (sg-Ctnnb1) or without (sg-NTC) deletion of Ctnnb1 were plated for colony forming assays at 10,000 cells per plate. Colonies were imaged and counted 10 days after plating (n=3). (f) Cell cycles phases in Lgr5fl/fl B-ALL cells with (Cre-ERT2) or without (ERT2) deletion of Lgr5 in combination with (sg-Ctnnb1) or without (sg-NTC) deletion of Ctnnb1 were studied by Edu incorporation and DAPI staining 3 days after 4-OHT treatment for Cre-mediated deletion of Lgr5. (g) Western blot analyses of β-catenin, Prdm1, Zap70 protein levels in bone marrow pre-B cells from Lgr5fl/fl or Ctnnb1ex3fl/+ mice in comparison to mature splenic B-cells from Lgr5fl/fl mice without (EV) and after activation of Cre. (h) Representative FACS staining of mature B cell populations in the spleens of Cd21Cre/+ Lgr5+/+ mice (no deletion of Lgr5), Mb1Cre/+ Lgr5fl/fl mice (early deletion of Lgr5 prior to pre-BCR checkpoint) and Cd21Cre/+ Lgr5fl/fl mice (late deletion of Lgr5, past the pre-BCR checkpoint) are shown.

bioRxiv preprint doi: https://doi.org/10.1101/2020.03.12.989277; this version posted March 13, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license. Figure 6: Pharmacological hyperactivation of β-catenin signaling represents a strategy to eradicate B-ALL. BCR-ABL1-driven Lgr5fl/fl (a) and Ctnnb1ex3fl/+ (b) B-ALL cells were transduced with 4-OHT-inducible Cre- ERT2 or ERT2. Gene expression changes following 4-OHT-treatment were studied by RNA-seq analysis (Fig. 3b). Gene set enrichment analysis showed depletion of MYC target gene expression defined by Hallmark- MYC Targets V1 but increases of IKZF1-upregulated gene sets. To elucidate why -catenin-activation results in transcriptional repression of Myc in pre-B cells, unlike other cell types, we transformed pre-B cells from Lgr5fl/fl x Ctnnb1flag/+ knockin mice that express FLAG-tagged -catenin. Upon Cre-mediated deletion of Lgr5, we identified binding partners by pull-down of FLAG-tagged -catenin and proteomic analyses (c). Pull-down with anti-IgG antibody was performed to normalize for non-specific binding. Proteins bound to β-catenin and IgG antibody were analyzed by mass spectrometry and are plotted based on significance and log2-fold enrichment over IgG (n=4). Proteins with the highest nuclear enrichment of binding to -catenin included Ikzf1 (Ikaros) and Ikzf3 (Aiolos) that function as transcriptional repressors of Myc in pre-B cells (highlighted in red). The -catenin-Ikzf1 interaction was confirmed by co-immunoprecipitation and Western blot (d), using antibodies to detect FLAG, β-catenin and Ikzf1 in whole cell lysates (input), proteins bound (elute) and not bound (flow-through) by IgG or β-catenin antibodies (d). Unlike colon cancer and CML, -catenin associates with Ikzf1 and Ikzf3 in pre-B cells, which could explain that -catenin accumulation (e.g. as the result of loss of Lgr5) results in transcriptional repression of Myc in pre-B cells as opposed to transcriptional activation in other cell types (e). B-ALL cells (red), CML (green) and colon cancer (gray) cell lines were treated with the GSK3 inhibitor LY2090314 at concentrations of (f) 0 up to 20 nM or (g) 0 up to 200 nM for 3 days. and cell viability was determined. All B-ALL cells expressed functional IKZF3. BV173 and MXP5 carry IKZF1 deletions. (h) B-ALL PDX and colon cancer (LOVO) cells were treated with the GSK3 inhibitor LY2090314 for 16 hours. Western blot analyses were performed for β-catenin, stable β-catenin lacking GSK3- phosphorylation (S33, S37, T41) and β-actin. (i) BV173 B-ALL cells were electroporated with Cas9-RNPs with non-targeting crRNAs (sg-NTC) or crRNAs targeting β-catenin (sg-CTNNB1). Single cell-derived colonies were generated, expanded in cell culture. Deletion of β-catenin was confirmed in two clones (1E4, 1C3) by Western blot (i). BV173 B-ALL cells (IKZF1-deletion) with (sg-CTNNB1) or without (sg-NTC) deletion of β- catenin were treated for 12 hours with LY2090314 (20 nM) and Western blot analyses for β-catenin and Myc levels were performed using -actin as loading control (j). B-ALL cells with (clones 1E4, 1C3; sg-CTNNB1) or without (clones 2C2, 2D9; sg-NTC) deletion of β-catenin were treated with LY2090314 for 3 days at the indicated concentrations and relative viability was determined by luminescence measurements (k). Colon cancer, B-ALL and CML cells were treated with LY2090314 at 10 nM for 12 hours and β-catenin and Myc levels were measured in these cells by Western blot, using -actin as loading control (l).

bioRxiv preprint doi: https://doi.org/10.1101/2020.03.12.989277; this version posted March 13, 2020. The copyright holder for this preprint Figure(which 1: wasIdentification not certified by ofpeer Lgr5 review) as is essential the author/funder, requirement who has granted for B-lymphopoiesis bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license.

Cre/+ fl/fl Mb1 a VH81-tg Inducible Blnk b Inducible Ig-HC d Lgr5 pre-B cells e Lgr5+/+ 10 14 Cre/+ P Mb1 =9.0E-04 Lgr5fl/+

=0.001 Cre/+ T2 Mb1

P ER Lgr5fl/fl 9 13 P P =0.001 =0.02 cells [% change] cells [% Splenic cells [%] Splenic cells 8 + Lgr5 mRNA levels [AU]Lgr5 mRNA GFP 12 Lgr5 surface levels [MFI] 0 0.5 1.5 5.0 0 24 Cre-ERT2 Induction [h] Induction [h] [days] Immature Marginal Follicular zone c Murine hematopoietic cells f Lgr5fl/fl, spleen Live cells B220+CD19+ B220+CD19+ 26 49 Mb1Cre/+ 57 Lgr5+/+ 33 5 34

1 10 Mb1Cre/+ 19 Lgr5fl/+

Lgr5 mRNA levels [AU]Lgr5 mRNA 56

s 18 -A -C C' -F lls lls lls lls lls lls r r r- r ge CLP F F Fr-D ce ce ce 59 , F , , Fr-E ce - cytesha LT-HSC -B B B lls, F lo ST-HSC o e- e e B- B1a NK-ce rop cre/+ r r re-B, r -ce GC B- 0 0 Mb1 P P P Pre-B, on CD8 TCD4 T-ce ac 0 atu B Granu Lgr5fl/fl m al z M 0 in Follicular B-cells Im g Naivear 0 B220 IgD M CD23 0

CD19 CD21 IgM

g Malignant Normal h k

12 P=0.025

10 8 6

LGR5 mRNA levels [AU] LGR5 mRNA n=31 n=153 LGR5 mRNA levels [AU] LGR5 mRNA LGR5 mRNA levels [AU] LGR5 mRNA MRD- MRD+

i lmCOG P9906 n=207 B-ALL COG P9906, n=43 Ph-like ALL 100 B-ALL B-cell lymphoma T-ALL AML 100 LGR5Low 80 80 LGR5Low LGR5 60 60 40 AKAP12 40 LGR5High 20 P= 0.008 20 P= 0.07 -actin LGR5High

0 Overall survival [%] 0 0 1 2 3 4 5 6

Relapse free survival [%] survival Relapse free 0 1 2 3 4 5 6 Time after diagnosis [years] Time after diagnosis [years]

j B-ALL B-cell lymphoma Myeloma T-ALL PTCL Myeloid BLQ5 TOM1 RAMOS RAJI U266 JURKAT HUT78 MV4-11 88 176 183 124 124 98

10,253 11,232

MXP5 MXP2 OCI-LY10 JEKO1 LP1 MOLT3 KARPAS299 CMLT1 2,379 123 80 177 120 135 110 LGR5 14,582

FSC bioRxiv preprint doi: https://doi.org/10.1101/2020.03.12.989277; this version posted March 13, 2020. The copyright holder for this preprint Figure(which 2: Lgr5 was not is certified essential by peer for review) B-ALL is the leukemia-initiation author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license. a Lgr5fl/fl BCR-ABL1 B-ALL b Lgr5fl/fl NRASG12D B-ALL ERT2 ERT2 cells [%] cells [%] + + GFP GFP

T2 Cre-ERT2 Cre-ER P P =0.0001 =0.0005 c ERT2 Cre-ERT2 d ERT2 Cre-ERT2 cells] cells] -4 -4 Colonies [10 Colonies [10 ERT2 Cre-ERT2 ERT2 Cre-ERT2 e Lgr5fl/fl BCR-ABL1 B-ALL 120,000 cells 6,000 cells 300 cells 17 days 23 days 36 days 17 days 23 days 36 days 17 days 23 days 36 days ERT2

Cre-ERT2

fgP=0.0006 P=8.2E-10 0.3K Cre-ERT2 0 100 6.0K Cre-ERT2 Cre-ERT2 LIC Cre-ERT2 1:40,055 ERT2 T2 LIC ER 1:1,031 80 0.3K ERT2 60 -1 40

20 -2 120K Cre-ERT2 0

0 20 40 60 80 100 120 [%] survival Disease-free 0 20 40 60 80 100 [days] Non-leukemic [log fraction] Non-leukemic Dose [cell number x103] 120K ERT2 6K ERT2 [cell number x 103]

h LGR5-PDX2 i Mouse BCR-ABL1 B-ALL cells 4 Primary Secondary Tertiary LGR5-MXP2 EV

LGR5-BLQ5 2

cells [fold change] Lgr5 + EV-MXP2 EV-PDX2

dsRed 1 EV-BLQ5 0 5 10 15 [days] Primary Secondary Tertiary j P =0.0001 0 LGR5 LIC LGR5 1:97 EV LIC EV 1:821 P =0.014 P =0.08 -1

-2 Colonies Colonies [per 10,000 cells]

Non-leukemic fraction [log] Non-leukemic fraction 0 0.4 0.8 1.2 EV Lgr5 EV Lgr5 EV Lgr5 Dose [cell number x 103] bioRxiv preprint doi: https://doi.org/10.1101/2020.03.12.989277; this version posted March 13, 2020. The copyright holder for this preprint Figure 3:(which LGR5 was notfunctions certified by as peer essential review) is thenegative author/funder, feedback who has regulator granted bioRxiv of  a-catenin license to display in B-ALL the preprint cells in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license. fl/fl fl/fl ex3fl/+ a Lgr5 B-ALL b Lgr5 B-ALL f Ctnnb1 B-ALL ERT2 Cre-ERT2 Axin2 [S244] ERT2 Cre-ERT2 ERT2 Cre-ERT2 Ctnnb1 [S675] Tigit Prdm1 326 332 Prdm1 Ccr2 Cxcr4 [S ,S ] Cd5 Akap12 Zap70 Zap70 Rbl2 [S1076,S1108] Ctla4 14 16 Dgka Pde8a [T ,S ] Ccr2 Ctla4 Rock2 Cd5 Cd244 Gnai2 [S282] Prkar1b Axin2 Il2ra Cblb Il2ra Dgka Cdkn2b n=83 n=354 n=242 Cdkn2b Akap12 Xbp1 Xbp1 Akap12 Cxcr4 E2f2 [S314] Axin2 Cd244 E2f2 Rag2 Myc Rag1 282 283 E2f2 Gnai1 Pax5 [T ,S ] Lgr5 Jund Myc [T79] Rag1 Myc n=50 n=61 Jun [S63] Rag2 n=119 [z-score] [z-score] [z-score] -1.5 0 1.5 -2 0 2 -2 0 2

c Lgr5fl/fl e Lgr5fl/fl Ctnnb1ex3fl/+ Lgr5fl/fl sg-Ctnnb1 g EV Cre EV Cre EV Cre EV Cre EV Cre EV Cre Gi -catenin

Akap12 Myc Lgr5fl/fl vs Ctnnb1ex3fl/+ B-ALL 9 Dgka -actin

Ccr2 -FC 6 2 Axin2 1A12 2F4 Cd5 Dgka Prdm1 Log Ctla4 -catenin- ex3fl/+ Cd244 33/37 41 Ctnnb1 B-ALL 3 Zap70

pS T h Axin2 ex3fl/+ Akap12 Cdkn2a -catenin NES = 2.4 .8 Il2ra 8 q < 0.001 0 -catenin-pS675 4 .4 E2f2 Ctnnb1 Myc Cdkn2a [Cre/EV] -3 0 0 Ranked genes -6 -30 3 6 9 fl/fl Prdm1 Lgr5 Log2-FC -actin i Lgr5fl/fl B-ALL j Ctnnb1ex3fl/+ B-ALL EV Cre EV Cre

106 484 46 449 CD5 CD5

1,796 8,447 2,147 7,522 Il2ra Il2ra

36 322 33 95 d LGR5-BirA WNT PKA Ccr2 8 Ccr2

-FC] LGR5 2 PGAM5 6 Cxcr4 Cxcr4 15,037 51,400 14,134 60,900 DVL2 PRKACB 148 308 99 116 4 AP2A2 Ctla4 AAK1 Ctla4 GRK2 BCL9L 2 RB1 AKAP12 74 666 353 4,126 DGKD CD244 ARRB2 MYCBP2 CD244 SUPB15 LGR5 vs EV [Log EV SUPB15 LGR5 vs 0 AKAP9 02468 CD19 CD19

PDX2 LGR5 vs EV [Log2-FC] bioRxiv preprint doi: https://doi.org/10.1101/2020.03.12.989277; this version posted March 13, 2020. The copyright holder for this preprint Figure 4:(which Lgr5-mediated was not certified bynegative peer review) regulation is the author/funder, of -catenin who has isgranted essential bioRxiv for a license B-lymphopoiesis to display the preprint and in B-ALLperpetuity. initiation It is made available under aCC-BY-NC-ND 4.0 International license.

a Ctnnb1ex3fl/+ B-ALL b Ctnnb1ex3fl/+ B-ALL [colonies/104 cells] c Ctnnb1ex3fl/+ B-ALL [cell cycle] ERT2 ERT2 59 P =4.0E0-4 T2 T2 ER Cre-ER G0/G1

40 1 S Cre- cells [%] cells T2 34 G /M + ER 2 GFP EdU T2 Cre-ER 59 5 EV Cre DAPI EV Cre

d Lgr5fl/fl B-ALL, rescue Ctnnb1-deletion e Lgr5fl/fl B-ALL, rescue Ctnnb1-deletion ERT2 sg-Ctnnb1 ERT2 Cre-ERT2 ERT2 sg-NTC

Cre-ERT2

sg-Ctnnb1 400 P =7.0E-05 cells [%] cells +

sg-NTC 300 GFP T2 Cre-ER 200 sg-NTC

100 sg-NTC sg-Ctnnb1 Ctnnb1 -catenin 0 sg- EV Cre EV Cre -actin sg-NTC sg-Ctnnb1

fl/fl fl/fl f Lgr5 B-ALL, rescue Ctnnb1-deletion Lgr5 B-ALL, Ctnnb1-deletion ERT2 Cre-ERT2 sg-NTC sg-Ctnnb1 sg-NTC 48 24

G0/G1 S 51 1 72 3

sg-Ctnnb1 G2/M 49 51 [% cell cycle phase] EdU 49 1 47 1 phase] cycle [% cell EV Cre EV Cre DAPI sg-NTC sg-Ctnnb1

h Live cells B220+CD19+ B220+CD19+ 49 60 Cd21cre/+ 62 Lgr5+/+ 23 10 g Pre-B cells Mature B-cells 29

Lgr5fl/fl Ctnnb1ex3fl/+ Lgr5fl/fl 1 8 Mb1cre/+ 12 fl/fl EV Cre EV Cre EV Cre 55 Lgr5

-catenin 29 72

Prdm1 50 69 cre/+ 70 CD21 Lgr5fl/fl Zap70 19

7 IgD CD23 B220 -actin 21

CD19 CD21 IgM bioRxiv preprint doi: https://doi.org/10.1101/2020.03.12.989277; this version posted March 13, 2020. The copyright holder for this preprint Figure 5:(which B-lymphoid was not certified malignancies by peer review) are is the exempt author/funder, from who oncogenic has granted activation bioRxiv a license of to-catenin display the signalingpreprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license.

a Activating CTNNB1 mutations [%] b Nuclear -catenin staining [%] Cancer types n 50 25201510 7550250 100 B-cell 2,375 B-cell Breast 2,683 Breast Esophageal 2,380 Esophageal Lung 6,730 Lung Prostate 3,305 Prostate Bladder 1,633 Bladder Melanoma 988 Melanoma Colorectal 10,707 Colorectal Glioma 5,635 Glioma Pancreatic 3,494 Pancreatic Ovarian 2,812 Ovarian Gastric 3,637 Gastric Renal 4,935 Renal Cervical 3,284 Cervical Liver 8,068 Liver

c Cell-type -catenin RPPA d Solid tumors B-cell malignancies 3 50 B-cell Other 25 0 [z-score] 2 0 [z-score]

-catenin RPPA RPPA -catenin -3 0 β -25 Dimension 2 Dimension [AU] -2 -6 -25 025 -25 025 Dimension 1 [AU]

e NSCLC Colon Melanoma B-ALL DLBCL MCL Burkitt’s HD Myeloma Nuclear - catenin

-tubulin

TBP Raji H82 Dld1 Lovo L428 ICN1 Z138 JJN3 U266 H146 H244 H249 H446 H524 H526 M285 M230 M229 BLQ5 MINO PDX2 SFO5 MXP2 MXP5 HT-29 KMH2 HBL-1 IAH8R JEKO1 Ramos SW620 SW480 HCT116 Gumbus SU-DHL2 SU-DHL4 OCI-LY10 Karpas422

f Normal colon Colon cancer Malignant melanoma NSCLC Breast cancer

Normal tonsil Mantle cell lymphoma Follicular lymphoma Diffuse B cell lymphoma Hodgkin’s lymphoma bioRxiv preprint doi: https://doi.org/10.1101/2020.03.12.989277; this version posted March 13, 2020. The copyright holder for this preprint Figure 6:(which Pharmacological was not certified by peer hyperactivation review) is the author/funder, of -catenin who has signaling granted bioRxiv represents a license toa displaystrategy the preprint to eradicate in perpetuity. B-ALL It is made available under aCC-BY-NC-ND 4.0 International license. a Lgr5fl/fl B-ALL b Ctnnb1ex3fl/+ B-ALL Hallmark Myc-targets V1 Ikzf1-upregulated genes Hallmark Myc-targets V1 Ikzf1-upregulated genes NES=-2.10 NES=1.60 NES=1.70

-fold] q-value=0.015 -fold] 2 q-value=0.01 2 q-value=0.02

NES=-2.45 q-value=0.004 Enrichment score Enrichment Cre-EV [Log Cre-EV [Log

Log2-fold ranked genes Log2-fold ranked genes Log2-fold ranked genes Log2-fold ranked genes

c Lgr5fl/fl x Ctnnb1Flag/+ d Lgr5fl/fl x Ctnnb1Flag/+ e Input Elute Flow-through FZD6 LGR5 FZD6 LGR5 Ctnnb1 9 Ikzf3 Apc ARRB2 ARRB2

p-value] Flag Lef1 Ikzf1 10 GSK3 DVL2 DVL2 6 Gsk3b Pgam5 β-catenin Axin1 -catenin PGAM5 GSK3 3 Ikzf1 Ikzf1 Ikzf1 Ikzf3 Ikzf3 -catenin 0 MYC MYC 052.5 7.5 10 12.5

Significance [-log Significance Positive pre-B cell selection Negative pre-B cell selection -catenin / IgG [log2] LIC-potential in B-ALL No LIC-potential in B-ALL

f LY2090314 g LY2090314 120 SW620 Colon LOVO Colon LOVO KCL22 CML SW480 90 SW480 CMLT1 CML SW620 JURLMK KCL22 60 JURLMK

BV173 B-ALL 30 MXP5

Relative luminescence [%] luminescence Relative PDX2 0 PDX2 B-ALL MXP2 [%] luminescence Relative [nmol/l] 0 50 100 150 200 [nmol/l]

h PDX2 MXP2 MXP5 LOVO j sg-NTC sg-CTNNB1 BV173 B-ALL - + - + - + - LY2090314 2C2 2D9 1E4 1C3 Clones - + - + - + - + LY2090314 -catenin -catenin -catenin Non- pS33/37T41 Myc -actin -actin k sg-CTNNB1 1E4 sg-CTNNB1 1C3 i BV173 B-ALL sg-NTC sg-CTNNB1 2C2 2D9 1E4 1C3 -catenin -actin sg-NTC 2C2 sg-NTC 2D9 Relative luminescence [%] luminescence Relative [nmol/l LY2090314] l Colon carcinoma B-ALL CML LOVOSW480 SW620 PDX2 MXP5 LAX2 IAH8R CMLT1 JURLMK1KCL22 K562 - + - + - + - + - + - + - + - + - + - + - + LY2090314

-catenin

Myc

-actin